How AI Is Changing Learning, Childhood and Human Connection

A podcast from Teachers College, Columbia University focusing on work in higher education that aims to improve our world.

Pursuing the Public Good is connected to a broader initiative at Teachers College focusing on efforts to tackle crucial challenges in Teacher Education, Mental Health and Wellness, Sustainability, and Digital Innovation.

Learn More about the Public Good Initiative

Thomas Bailey:
Welcome to Pursuing the Public Good. I'm Tom Bailey, President of Teachers College. This podcast is part of something I care deeply about, and is at the heart of what we do here at TC. It's called The Public Good Initiative, and it's a commitment that our researchers, scholars, and students will collaborate across disciplines and across communities to confront the most pressing issues of our time in education, health, and psychology. In this episode, we hone in on one of the most consequential questions we face in education today. What does artificial intelligence mean for how we teach, how we learn, and how we prepare young people for a world that is changing faster than any of us can fully anticipate?
Technology doesn't wait for us to catch up. Our schools, our families, and our students need guidance today. We must ask hard questions like, what is lost and what might be gained when a student turns to a chatbot instead of wrestling with difficult ideas on their own? What does AI do to a child's developing sense of self? In the conversation you're about to hear, journalist Jill Barshay of The Hechinger Report sits down with Lalitha Vasudevan, Vice Dean for Digital Innovation and Managing Director of The Digital Futures Institute right here at Teachers College. Lalitha not only researches these questions, but advises educators on how to navigate them.
They're joined by Amanda Lenhart, a Senior Fellow at the Joan Ganz Cooney Center, who has spent years studying how technology shapes child and adolescent development. Together, they explore what AI is actually doing in the classrooms today. From the AI already used in schools, to the very real risk that students are outsourcing, just not homework, but the challenge that makes learning meaningful. They talk about what parents should do at home, how schools should think about policy, and why the procrastinating student with an essay due tomorrow is, in many ways, the perfect lens through which to view this whole challenge.

Jill Barshay:
Thank you for joining us on season two of Pursuing the Public Good. I'm Jill Barshay. I'm a journalist at The Hechinger Report, where I've been covering whether the new uses of artificial intelligence and education are really helping children learn. I'm joined by Lalitha Vasudevan.

Lalitha Vasudevan:
Hi, Jill.

Jill Barshay:
We also have Amanda Lenhart.

Amanda Lenhart:
Hi, Jill.

Jill Barshay:
There's been a lot of discussion, Amanda, about how smartphones and screens have reshaped childhood and adolescence. From your research, what feels most different about being a child or teenager today compared with, say, 30 years ago?

Amanda Lenhart:
It's a great question. So many things. Certainly, I think the big ones that come to mind are the constant connectivity that comes from a smartphone in your pocket, which the vast majority of adolescents over 13 have, from the social platforms where adolescents spend a lot of their time and have a lot of their peer interactions. Now, bringing in generative artificial intelligence and all the things that you can do and experience and that can happen with generative AI. All of those things, I think, are really reshaping both how you need to present yourself to the world in particular, how you go about learning things, and also, how you expect to be connected to other people in your life.
That you always have that in your pocket to text or phone a friend for an opinion about something, get people together ambiently. I think there's a lot of things that have changed from 30 years ago to today.

Jill Barshay:
Let's dig into the AI part of it because that's the topic of this episode. How are you already seeing AI affect teens' social lives or inner worlds? Maybe you can give an example of a really interesting or creative use of AI that you're seeing teens engage in.

Amanda Lenhart:
Yeah, I think it's really important to remember that this is a technology we've had, generative AI in particular, for really about three and a half years. We are very early days. I think some of the deeper questions about how it's changing our internal worlds are ones that I think are continuing to evolve, but I think we can talk a lot about what is happening with adolescents. We do see, and in fact, even I have some research that I did almost two and a half years ago now, so right at the beginning, where we asked adolescents about things they thought adults didn't understand about what they were doing with generative AI.
Even then, we heard about some really fascinating, but also hair-raising things that young people are doing, including making music, making videos. Things that, again, today don't even sound that surprising given how rapidly the technology has evolved. But also, things like impersonating other people for good reasons or bad reasons, using it for sexual talk and play with another character online. Even back then, adolescents were already beginning to push the envelope into some of both creative uses, academic uses, but also transgressive uses. I think there's a lot of ways in which this is really beginning to infiltrate adolescents' lives.
But that said, not every teen understands the technology super well, necessarily is like, "I want to get away from a digital native perspective." They're not always on top of what it's really doing, but there are a lot of ways in which it is now very present in their lives.

Jill Barshay:
Let's bring Lalitha into the conversation, particularly on the academic uses of AI. What kinds of tools are already showing up in classrooms and how do you see them being used beyond cheating and having it write the paper for you?

Lalitha Vasudevan:
Yeah, this was the concern when generative AI first exploded on the scene. As Amanda said, we're really early in this phase of the AI story. Initially, even here on our campus, but also in schools around the country, there was a desire to rush to ban because it was such an unknown. Some of the early uses were just, oh my God, look, it can write a whole paper for me. One of the things that's been interesting, and I've been talking with some schools and some teachers, is they are trying to navigate this tension between how much access to give, what is our responsibility in terms of turning on or off the AI capabilities that are built into tools that schools are already using?
This is the big thing. It's less about introducing novel AI tools into classrooms sometimes. It's about being aware of what's already baked in to the technology that schools are using.

Jill Barshay:
Just give us an example, what AI has baked in?

Lalitha Vasudevan:
Google turned on Gemini for everybody to use. Up until about a year ago, they had said, "We're going to leave it off for kids under 13." Then I think that rule or guardrail went away. Now it's at the school level to decide whether to turn things on or not. But even as people are using Google Docs, a lot of schools that are partnered with Google they have Google Docs, they've developed their curriculum around it. There's a collaboration that happens.

Jill Barshay:
That's how you turn in your homework.

Lalitha Vasudevan:
It's how you turn in your homework, Google Classroom. But some schools, the Google Suite has a capability to just have Gemini turned on. A little question will pop up. Do you want some help making this look better? Do you want some help editing this? Now, one of the things that the teachers we've spoken with are wrestling with is, what is their responsibility in terms of preparing young people for a work world where a lot of these tools are being rapidly adopted? Even as these small tech companies might be folding very quickly and new products are developing, so you're constantly having to learn to use different products. Where does that leave schools in terms of being able to be steady preparers of thinkers in the world rather than just adopters of new technology?

Jill Barshay:
Well, that leads into my next question for you, Lalitha. You've been working closely with teachers and school leaders on how they can adjust to AI. What's the most important advice you're giving them right now?

Lalitha Vasudevan:
Yeah, I think one of the things that schools tend to do is they want to put policies into place. We're going to ban this, we're going to allow this. New York City right now is navigating this at a district level. They have their stoplight policy. What's green is go. What's yellow is caution. What's red is verboten.

Jill Barshay:
If I could just interrupt for our audience.

Lalitha Vasudevan:
Please.

Jill Barshay:
The idea is that teachers can have three ways of dealing with AI. One, you can use it all the time. Two, banned, never use it. Yellow is some mishy-mash in between.

Lalitha Vasudevan:
Absolutely. Even that, I think as soon as you make something a policy, it's really hard to walk back from that. One of the things we're talking to schools about is thinking about the difference between policy and guidelines. Now, this seems like a tiny semantic difference.

Jill Barshay:
It sounds like something academics would distinguish.

Lalitha Vasudevan:
It is, it is. I say that because the technology and what we know about it is changing so rapidly that when you put something in policy, typically, policies have to be voted on or vetted and gone through processes. But if you offer guidelines that says, "You know what? We're going to check back in a month. Is this working?" That I think allows two things to happen. One, it allows you not to be in a fixed position, even as the technological world is rapidly changing. But two, it allows more voices to help shape the process, including young people's voices. Yeah, Amanda, you want to jump in here?

Amanda Lenhart:
Yeah. I would say this is so helpful. In some of the work that I've done, we talk to parents and kids about some of these exact questions, and they're totally looking for the guidance. They're totally looking for things from teachers and schools, both parents and students, about what's the right thing to do. What are the definitions of the yellow in the middle? What are the things that we should be doing? It's not always easy. Just a quick story, I ran a youth panel at my PTA on a virtual meeting to talk about generative AI. We also had a member of the school staff on the panel as well, a counselor. We got into it a little bit.
The students wanted to bring up what they felt was the unfairness of teachers using the AI to make some of the products that they saw and the worksheets and the tests in school that were very obvious to them created by AI. But then they were told as middle schoolers that they were not allowed to use it. They thought this was grossly unfair. I would say the friction and the challenge is, we create these policies and then the implementation of them with actual children is where it gets messy on the ground.

Jill Barshay:
What I hear both of you saying is, let's not put anything in stone right now, let's try something and get some feedback and reassess?

Lalitha Vasudevan:
I think so. Part of the consequence of putting things in stone is that schools need to enforce it. If you have a policy that says you can't use it, a student uses something that falls under the red zone, the school has no recourse but to act on it. I don't know that we want to start punishing kids for something we're just figuring out as a field. That's where the real consequences come in. Now, clearly, you want to work with teens, young people, middle schoolers, high schoolers, college students to talk about ethical appropriate uses. Not only some of the uses that Amanda was bringing up about things that may be not great for their own wellbeing or seeing ways that students are saying, "Oh, a friend told me that I could just ask this tool questions about advice or about health issues."
You want to really be scaffolding that because that can have real life consequences. But you also want to, I think, not box yourself in as a school leader that says, "Oh, well, the policy says this. I have to act on this and now you're suspended." I think there's a little bit more nuance that we need.

Jill Barshay:
Let's circle to the enforcement a little bit later because I have a bunch of questions about that. But you've been doing some very interesting research with teachers on their attitudes about AI, especially with writing. I was wondering if you could give us a preview of what you're seeing so far.

Lalitha Vasudevan:
Yeah, we're in the middle of this research, so I'll give some preliminary things. We've done some background literature review and we see the typical, not quite spectrum but split, teachers who are completely rejecting. AI has no place in writing because writing is thinking, writing is becoming, writing is expression. Then we have some people who are navigating this a little bit and the workforce argument. It's in the world. We need to help teachers, our students understand what this is. We as teachers need to understand what this is. Let's think about what the writing process is, where we can see these tools, and not just one tool, but this broader technology being useful.

Jill Barshay:
When you talk about not just one tool, but there's different steps to the writing process like research, brainstorming, outlining. Maybe there's a role for it in some of these steps.

Lalitha Vasudevan:
That's exactly right. Then the third piece is where people are finding those ways to integrate. One area where there's been some research done, largely outside of the US that we have found, is looking at the ways that teachers are using AI tools for feedback. How are they offering feedback for student writing? They might put in a rubric, they might custom train a LLM to give the kind of feedback they might give. I was talking to a colleague and it was interesting because we were talking about how that might be editing feedback, but it's not really feedback on writing in the way that a teacher who knows her students can give feedback on the student as a writer.
This is something that is coming up as well. What is my job as a teacher of writing? As a teacher who teaches writing, who's trying to cultivate writers in my class and really trying to help emerging writers think about what does it mean to write?

Jill Barshay:
Right. It's funny, I was looking at a research study a couple years ago now that compared AI writing feedback to the feedback of expert writing teachers, and I was shocked by how good the feedback was. I often preferred the specificity and the help of the AI writing feedback. Sometimes the humans would be like, needs clarity, and that's not so clear, please.

Amanda Lenhart:
Just to jump in here too, I think that this brings up bigger questions about expertise and learning. I think writing is a perfect microcosm of a lot of the challenges that we have with the ways in which AI is showing up in our academic lives. How do we get to that level of expertise that you have as a mid-career or even a more experienced professional if you never have the chance to go through the process of doing it yourself? I think that's a thing that if we're talking about and we're focusing on workforce issues, well, how do we make sure that our young people, our writers, our engineers, our coders, our whoever has enough experience to look at other things and be like, "That's not good for these reasons." I think that is one of the big things that I worry about. I think the writing example is a beautiful focus on that broader question.

Jill Barshay:
Let's go back to the traffic lights. I'm fascinated with red light, green light, yellow light. I'm so curious to hear both from Amanda and from Lalitha. If you were teaching in a classroom, what color light would you have? How much AI would you be having in your classroom today?

Amanda Lenhart:
I think that answer would really depend on the age of the students I was teaching. If I were teaching younger kids, I probably wouldn't have it at all, honestly, because I think some of the real foundational skills need to get laid before AI comes in. Now, of course, they could be using it at home, they could be using it with parents, they could be using it with friends. There's lots of ways in which they might be exposed to it and using it. As you get older, I think I would integrate it in smaller ways, again, in an assistive fashion where it's okay to run a draft that you've written through a document to get feedback on it. But I don't want to see you wholesale write a thing.
I want to see your ideas. I want you to generate them. I think I would have more of an assistive focus. As you got older and older, as we move into high school, I would, again, loosen the reins a little bit. But in general, I think I am a little bit more of a keep it out while you're developing your skills so that you can bring it in a little bit later and start to really learn how to use it well.

Jill Barshay:
You're a red light, yellow light woman?

Amanda Lenhart:
Maybe red light, yellow light until we get up into... Yeah, I think probably I am, until we get out of the K-12. I think I'm probably not a full green lighter.

Jill Barshay:
Yeah, Lalitha?

Lalitha Vasudevan:
Yeah, that really resonates. I think part of what would inform the decision would be also what we're talking about when we're talking about AI. A lot of times we're talking about chatbots because they're in the atmosphere. But we know certainly in graduate school, people are thinking about and utilizing tools that help with processing data, doing large data sets, and querying them for patterns. The caution, of course, with all of this is awareness of data privacy, doing any kind of AI engagement alongside real robust curriculum around, what are we doing when we're using this? Really inviting students and learners and teachers to ask, "What do I need to be able to do, and what's the broadest palette of resources I can use for this?"
Now in my previous teaching, and both in working with middle schoolers and high schoolers and graduate students, one of the things I have always loved to do is have time when we just say, "We actually don't have a goal, but let's see what this technology can do." I think there's room for that as well, more in the older years, I think. In part because, and just to echo what Amanda said, when you have a five and six-year-old, their understanding is about the world and their own imagination. I would hope we have that as capacious as possible. We keep the boundaries of that as wide as possible. I want a kid coming in, when you ask them what do you want to be when you grow up, at five and they say, "I want to be a fish."
You're like, "Yup, that's totally possible. You might turn into a fish." We are not going to get that if we start saying, well, here's the answer, always, rather than, what are the questions? I don't know where I sit on the, I was going to say red, white and blue, on the green, yellow, and red scale but it depends is always going to be the answer. I would hope that there's more people engaged in the tinkering of these things, while we help shape what might be the real uses for them. One of the things that teachers also bring up is, hey, there are so many tools my school is always pushing on me to help me with lesson planning, to help me with student feedback, but that's what I do as a teacher.
Can I give them feedback and say, "Actually, lesson planning, I got that. But you know what would be great is if you help me prepare these reports that I have to submit." Even tweaking what we support learners in this case to do would be really helpful.

Jill Barshay:
Right. We're talking a bit about guidance for teachers. Amanda, I'd like to switch to guidance for parents. I personally can use this guidance right now. What kind of boundaries should we set for our children with AI at home?

Amanda Lenhart:
This is such a really tough question, and this is a caveat to even the side card use in academic environments, is right now the large language models, which is the big thing that most people use, as Lalitha was saying, there's visible uses of AI and invisible uses. When it just shows up in your search or you use it in your Adobe product, you don't even necessarily know that it's there except for it's a little sparkly icon. But when you go to a chatbot, you know that you're going there and you acknowledge that you're going to generative AI. The challenge with those large language models is that they are large and they do everything.
I think one of the real challenges that I see is that while they have all this potential for, I need to vibe code myself an app to help with my homework on this, I need to be able to plan a trip with my family, I want to learn a new skill and I want to check this other thing. Super powerful, but the challenge is that what has ended up happening is because they also are relational. These chatbots in particular are designed to feel human-like, and they're designed to play on our needs for human attachment, and that's how they keep us going. We went from the attention economy of a social media platform and we're moving into the attachment economy.
I think what's really challenging is that's not a great thing for kids. For people who are learning how to become good socialized humans with other humans, an endlessly positive, endlessly sycophantic chatbot, even if it's helping you with your math homework, may not be optimal for you as somebody who's learning how to develop relationships with other people. The other challenge is that so many people, young people, use it and they swing back and forth. You start out asking about homework and then you're like, I just got dumped, I'm feeling sad. I'm going to ask a question on the chatbot. I don't really want to talk about it with anybody else.
Nobody wants to hear me moan anymore. That's when you swing into these more relational, effective, companion-like conversations. I think that's where I have my concerns.

Jill Barshay:
Right, because you might not be developing. You could have called a girlfriend and developed a stronger bond discussing your failed relationship.

Amanda Lenhart:
Yeah, or with a peer friend, with a parent, with somebody else. Then the ways in that often leads into more companion-like relationship is that it becomes my buddy, it becomes my friend, it becomes my AI girlfriend/boyfriend, which is real. We've heard that for two years. People are really using it. They get better and better as they make fewer mistakes, as they become more human-like. They become, I think, more challenging in that regard, and people like them because they are like that. I personally struggle with this in terms of making clear the kinds of uses that you're more comfortable with. If my child had questions about homework, I would prefer that they use a purpose-built AI tutoring platform as opposed to an open large language model.
We talk about the ways in which you use that with your child. Okay, hey, we're planning a trip. We're going to work on it together. We'll sit down. We do it side by side. That becomes much harder as your adolescent gets older. Again, for those who are really interested in being creative and creating with the chatbots, I don't want to limit that, but I have real reservations about allowing young adolescents, in particular 14 and under, using the platforms unfettered with no supervision. There are now ways to age identify your child and they go into a more gated role on the LLMs. But to be clear, there is a industry-wide problem with guardrails becoming softened.
The various security features put into place to keep adolescents safe and children safe on these, if you are on them for a very long time, they fall away. Nobody really knows why. They're working on getting that better, but they're not better yet. I would say that for me is also a real pause, even in the context of perhaps a more age-appropriate experience, which to be clear, the chatbots already know how old you are. It takes them a couple of weeks to figure it out, but you will have it. You don't need to identify in any way. They will just figure it out from your conversations for the most part. Again, I think we're in an interesting space.
But I personally have a lot of reservations. I would say parents that are worried about it, I would have real concerns about unfettered use by children without supervision under the high school age.

Jill Barshay:
Right, a lot to think about. Let's take a specific use case. If other people's daughters are anything like mine, they occasionally procrastinate. I think about how realistic is it that we can ask students or teens to use AI responsibly when Claude or Chat can instantly generate an essay on their Jane Austin novel the night before it's due. I'm curious for both of you how you talk to teens about this.

Lalitha Vasudevan:
Yeah. Look, as a card-carrying member of the procrastination club, I appreciate this question very much because when that first boom of GenAI tools in ChatGPT came on the scene about three and a half years ago, people were rushing to put things in place. One of the things that came back, certainly at the graduate school level, but I know in other schools was, okay, we're going to do everything oral exams. We're going to do blue book exams. We're going to get rid of the opportunity for you to do anything at home. Then I think certainly as our institute started working more with our faculty here, but also teachers that we work with, we started inviting them to ask why they were giving the assignments they were given.
For a long time, curriculum in schools has just followed what people have always done. I don't know if silver lining's the right phrasing, but one of the things that AI has put front and center is a need to really be intentional and purposeful about what we're asking students to do in our classrooms, which should always be the case. But the reason we're having teachers really think about this is because of this instant production. Not only instant production, but this opportunity to engage with a tool basically that's going to tell you you're amazing, that you have these great ideas. It's going to produce something that you're going to say, "Well, the thing said it was amazing." Then what happens if the teacher says, "What is this?"
There's that element. But we also, I think, have to work with teachers to have open conversations with students that says, not just a paper is due on this date, but I want to see your outline on this date. I want to see your first draft on this. It's really focusing more on explicating process and being partners with the students. That's certainly in middle high school. I also agree with you. It is unfair to assume individual responsibility all the time. I would love it if tech companies would really get away from the black boxing that is happening with a lot of the generative AI tools that students and people are using and said, this is actually how this technology works.
It allows responsibility to be shared a little bit more rather than individual use responsibility, which is often what policies are driven by. But it also, I think, helps to refocus the energy around scaffolding engagement in thoughtful ways and allowing the user of these tools to ask, what am I trying to do, rather than, what can this produce for me right now?

Jill Barshay:
Amanda, I'd like to bring you in. Once this has happened, once a teenager maybe has used a little bit too much AI help, what are your thoughts on how to address that at home?

Amanda Lenhart:
Yeah, I think these are really challenging questions. Even I just want to pick up on Lalitha's point about engagement. My sense is, in talking to young people, that generative AI is something that you use when you're not engaged. It's something that you use when you're not motivated, you don't care. It's for the class that you don't like, the teacher who you don't feel seen by. When you don't understand what education's for and why you might want to learn these skills. I think there's work to be done in terms of buy-in for a subset of young people who frankly are not particularly interested in the work that they're doing, and that I think is a super big challenge. It's been a challenge forever.
But I think expecting young people who have this free tool that can do a pretty darn good essay at the tips of their fingers for a class that they're not that engaged in, that they're not going to take that choice. If you don't, for some people you're a chump. Why wouldn't I do this? Why wouldn't I? People do it the hard way, why would you? I'm never going to do this in my future life. Again, explaining why you might want to do this, why you need to learn how to write this, why you need to know how to do this type of thing. Why you need to be able to write a strong letter, why you need to be able to convey certain things to yourself the value of the things that you're learning.
I think that's a really difficult underpinning question here. As for what to do in terms of like, you did the thing and I'm not pleased about it. Again, this is I think a strong parenting moment where you can decide as a parent. If it were me, I would out my child to their teacher and I would have them face the consequences of it.

Jill Barshay:
You're tough.

Amanda Lenhart:
I would have them rewrite it. Yeah, no, I would have them redo it because I don't think it's... I think for me, and we talked about this a lot. I have four children and my youngest is 13. My oldest are teachers themselves. Some of them are teachers. We talk about these things a lot in all sort of family dinners when we have them, and we talk a lot about the value of learning to do the thing. We impress that upon all of our kids as to why, and that big broader family conversation happens. But that said, I think there are some parent-child relationships in adolescents that are very challenging and where that may not go over very well. Then, again, you're set with the idea of, how much do you want to limit access?
If that's a thing that's important to you, again, I'm more of the let's talk about it and talk it through and explain why. But for some families, that's not a way they want to go. In that case, you may want to have time limits, bans, supervision in stronger ways. Again, not usually my practice, but again, I acknowledge that there are families where that is their preferred method or where they are based on their relationship with their child.

Jill Barshay:
Lalitha, earlier in our discussion, you were talking about enforcement. I think this is a moment you discover that the student has not done his or her writing work. Yes, maybe there are preventative things you could have done beforehand, restructuring the assignments and also chunking them into micro morsels along the way. But once you're confronted with the infraction, suspending sounds draconian, but what do you do?

Lalitha Vasudevan:
I'm going to build on something Amanda just said, about inviting a conversation. Simply starting by asking, tell me why. Tell me why you did this. Were you feeling stressed? Was there not enough time? I would also have a conversation about, what do you know about AI? Are you using it in other places? There's been some research that people have looked at, specifically looking at teenagers' uses of social media filters that were all AI generated, and you could create filters of yourself looking like a nurse, an astronaut. I don't know what other careers there are apparently, but other things. When asked about that, teens were saying, "This is motivating for me."
Trying to understand the repertoire of teens' tech use, of their AI use, certainly understanding teens' tech use beyond what is visible, is something that I think a lot of researchers, including Amanda, have tried to impress upon educators. Teens are doing a lot of really creative, thoughtful things in the world. Where is their space in a school curriculum, in a classroom setting to go beyond what you can see within the confines of your curriculum? I would invite teachers to really think about students as partners in learning at this particularly really important moment, in what is going to be a very long AI story in this country.
We have an opportunity to not just accept a narrative that's being given to teachers and students, but instead help craft that in a different kind of way. Just starting with, tell me about this. What's going on here?

Amanda Lenhart:
Yeah, I want to stand up and clap. That was amazing. Absolutely.

Jill Barshay:
Let's wrap up this conversation with a fear and a hope. I'd like to ask both of you, what is your biggest fear of how AI might play out with teens and schools? What is your greatest hope and feeling of optimism of how AI could make this world a better place? I'll start with you, Amanda.

Amanda Lenhart:
Yeah, I think my fear is that we continue on the paths that we're on right now. That we just move forward focusing exclusively, frankly, on preparation and not on the whole child and what this is bringing to the whole child's life experience. From their social and emotional development, as well as their skill development, as well as their preparation for their future. I think the more we can take a holistic view, I think the better off we'll be. I think I would be fearful if we throw up our hands as well and we say that, "Okay, we're just not going to teach this anymore. Everybody's going to do it on AI, that's fine." We're just going to work towards management as opposed to skill development.
I think that would be a real loss for us as citizens of the country for developing people who can think in smart ways and write in smart ways, since we do communicate so much through writing. I think that would be a real loss. I think my hope is that, again, we focus on the whole child. We focus on the whole person, and we figure out ways to frankly silo off the different functions of AI so that we create great places to get tutoring. We create bespoke writing assistants. We create places where you can go and get social and emotional support. But that they're all specific and they have the unique guardrails that each of those different types of uses needs.
And that we can use them in ways that are beneficial while excluding some of the really more problematic and concerning relational and potential harms that we're seeing emerge.

Jill Barshay:
Lalitha?

Lalitha Vasudevan:
You know the movie Hidden Figures, which is based on Katherine Johnson's book and her life? I think about that shift that happened from human computers, people who were computing, to the use of computers as we know them now. One of the things that I think about from that story mostly is, given the sociopolitical tensions of the time, how hard it was for Black women to be accepted into that world of NASA, and how much would've gotten lost if Katherine Johnson hadn't been part of that mission that eventually got us to space? If you scale that out, that we didn't allow our society to grow because of these sociopolitical forces.
I feel there's a corollary that we could look to now, which is the rules are being set by very few companies, entities. Again, to echo what Amanda just said, a fear I would have is that we are complacent to that. That we say, "This is the way life is. This is the way the rules are. We're not going to regulate, and therefore we just have to adopt the tools that exist." If we let that complacency sink into our schools and we allow future generations to only see themselves in the ways that these tools are reflecting back to them, then we will have shut off the broadest sense of imagination that allows a society to evolve and develop. We need scientists, we need teachers, we need artists.
If we're going to seed all that ground to algorithms, then we are going to shrink ourselves as a global society. Now, my hope is that this moment invites us to invest more in those fundamental things that make us human. The arts, the humanities, the empathic ways that we relate to one another, so that we can get to what Amanda was saying, which is harnessing the computing capacities of this technology to be able to create really specifically targeted scaffolding tutors for learning math, supporting writing, mental health support, because the technological capacity is fascinating. What we're currently doing with it is boring. If this thing can do all these things, why are there still diseases? 
Why are there still all of these social ills? Let's use the computing capacity in generative social ways, and I think we need really strong humans to help guide that.

Jill Barshay:
That's a nice ending, that we still need humans in the loop somewhere.

Lalitha Vasudevan:
We can only need humans.

Jill Barshay:
Well, thank you. I want to thank you both for this illuminating conversation that I'm going to take home to my family tonight. This was Lalitha Vasudevan, Vice Dean for Digital Innovation and Managing Director of the Digital Futures Institute at Teachers College. And we had with us Amanda Lenhart, senior fellow at the Joan Ganz Cooney Center. I want to thank you both for your time today. I hope that your greatest hopes for AI future come to pass.

Lalitha Vasudevan:
Thanks, Jill.

Amanda Lenhart:
Thank you.

Thomas Bailey:
This conversation leaves me with a sense of cautious optimism. The risks are real, but so is the possibility. If we insist on keeping students, teachers, and families at the center of this conversation rather than on the sidelines, we can shape how this technology develops. Artificial intelligence is one of the defining challenges of our time, and how we respond in education will matter enormously for the kind of society we become. I hope this conversation sparks something for you, and thank you for listening to Pursuing the Public Good. Please follow us wherever you listen. If you like the episode, leave us a review and share it with a friend. Thank you for listening.